Download presentation

Presentation is loading. Please wait.

Published byEthan McDonald Modified over 2 years ago

1
Data Assimilation Training Course, Reading, 5-14 May 2010 Tangent linear and adjoint models for variational data assimilation Angela Benedetti with contributions from: Marta Janisková, Yannick Tremolet, Philippe Lopez, Lars Isaksen, and Gabor Radnoti

2
Data Assimilation Training Course, Reading, 5-14 May 2010 Introduction 4D-Var is based on minimization of a cost function which measures the distance between the model with respect to the observations and with respect to the background state The cost function and its gradient are needed in the minimization. The tangent linear model provides a computationally efficient (although approximate) way to calculate the model trajectory, and from it the cost function. The adjoint model is a very efficient tool to compute the gradient of the cost function. Overview: –Introduction to 4D-Var –General definitions of Tangent Linear and Adjoint models and why they are extremely useful in variational assimilation – Writing TL and AD models –Testing them –Automatic differentiation software (more on this in the afternoon)

3
Data Assimilation Training Course, Reading, 5-14 May D-Var In 4D-Var the cost function can be expressed as follows: B background error covariance matrix, R observation error covariance matrix (instrumental + interpolation + observation operator error), H observation operator (model space observation space), M forward nonlinear forecast model (time evolution of the model state). H T = adjoint of observation operator and M T = adjoint of forecast model.

4
Data Assimilation Training Course, Reading, 5-14 May 2010 Incremental 4D-Var at ECMWF The gradient of the cost function to be minimized is: In incremental 4D-Var, the cost function is minimized in terms of increments: with the model state defined at any time t i as: 4D-Var cost function can then be approximated to the first order by: where is the so-called departure. and are the tangent linear models which are used in the computations of incremental updates during the minimization (iterative procedure). and are the adjoint models which are used to obtain the gradient of the cost function with respect to the initial condition. at t=0) is the trajectory around which the linearization is performed (

5
Data Assimilation Training Course, Reading, 5-14 May 2010 where is the linearised version of about and are the departures from observations. Details on linearisation In the first order approximation, a perturbation of the control variable (initial condition) evolves according to the tangent linear model: where i is the time-step. The perturbation of the cost function around the initial state is:

6
Data Assimilation Training Course, Reading, 5-14 May 2010 The gradient of the cost function with respect to is given by: Details of the linearisation (cnt.) The optimal initial perturbation is obtained by finding the value of for which: The gradient of the cost function with respect to the initial condition is provided by the adjoint solution at time t=0. remembering that

7
Data Assimilation Training Course, Reading, 5-14 May 2010 For any linear operator there exist an adjoint operator such as: Definition of adjoint operator where is an inner scalar product and x, y are vectors (or functions) of the space where this product is defined. It can be shown that for the inner product defined in the Euclidean space : We will now show that the gradient of the cost function at time t=0 is provided by the solution of the adjoint equations at the same time:

8
Data Assimilation Training Course, Reading, 5-14 May 2010 Usually the initial guess is chosen to be equal to the background so that the initial perturbation The gradient of the cost function is hence simplified as: Adjoint solution We choose the solution of the adjoint system as follows: We then substitute progressively the solution into the expression for

9
Data Assimilation Training Course, Reading, 5-14 May 2010 Adjoint solution (cnt.) Finally, regrouping and remembering that and that and we obtain the following equality: The gradient of the cost function with respect to the control variable (initial condition) is obtained by a backward integration of the adjoint model.

10
Data Assimilation Training Course, Reading, 5-14 May 2010 Iterative steps in the 4D-Var Algorithm 1.Integrate forward model gives. 2.Integrate adjoint model backwards gives. 3.If then stop. 4.Compute descent direction (Newton, CG, …). 5.Compute step size : 6.Update initial condition:

11
Data Assimilation Training Course, Reading, 5-14 May 2010 Finding the minimum of cost function J iterative minimization procedure cost function J model variable x 2 model variable x 1 J(xb)J(xb) J mini

12
Data Assimilation Training Course, Reading, 5-14 May 2010 An analysis cycle in 4D-Var 1 st ifstraj: Non-linear model is used to compute the high-res trajectory (T1279 operational, 12-h forecast) High-res departures are computed at exact obs time and location Trajectory is interpolated at low res (T159) 1 st ifsmin (70 iterations): Iterative minimization at T159 resolution Tangent linear with simplified physics to calculate the increments The Adjoint is used to compute the gradient of the cost function with respect to the departure in initial condition Analysis increment at initial time is interpolated back linearly from low-res to high-res and it provides a new initial state for the 2 nd trajectory run 2 nd ifstraj: repeat 1 st ifstraj and interpolates at T255 resolution 2 nd ifsmin (30 iterations): repeat 1 st ifsmin at T255 Last ifstraj: Uses updated initial condition to run another 12-h forecast and stores analysis departures in the Observational Data Base (ODB) 2 minimizations in the old configuration Now 3 minimizations are operational!

13
Data Assimilation Training Course, Reading, 5-14 May 2010 Brief summary on TL and AD models

14
Data Assimilation Training Course, Reading, 5-14 May 2010 Simple example of adjoint writing

15
Data Assimilation Training Course, Reading, 5-14 May 2010 As an alternative to the matrix method, adjoint coding can be carried out using a line-by-line approach. Simple example of adjoint writing (cnt.) (often the adjoint variables are indicated in literature with an asterisk)

16
Data Assimilation Training Course, Reading, 5-14 May 2010 More practical examples on adjont coding: the Lorenz model where is the time, the Prandtl number, the Rayleigh number, the aspect ratio, the intensity of convection, the maximum temperature difference and the stratification change due to convection (see references).

17
Data Assimilation Training Course, Reading, 5-14 May 2010 The linear code in Fortran Linearize each line of the code one by one: dxdt(1) =-p*x(1) +p*x(2) :Nonlinear statement (1)dxdt_tl(1)=-p*x_tl(1)+p*x_tl(2) :Tangent linear dxdt(2) = x(1)*(r-x(3)) -x(2) (2)dxdt_tl(2)=x_tl(1)*(r-x(3)) -x(1)*x_tl(3) -x_tl(2) …etc If we drop the _tl subscripts and replace the trajectory x, with x5 (as it is per convention in the ECMWF code), the tangent linear equations become: (1) dxdt(1)=-p*x(1)+p*x(2) (2) dxdt(2)=x(1)*(r-x5(3))-x5(1)*x(3)-x(2) … Similarly, the adjoint variables in the IFS are indicated without any subscripts (it saves time when writing tangent linear and adjoint codes).

18
Data Assimilation Training Course, Reading, 5-14 May 2010 Trajectory The trajectory has to be available. It can be: saved which costs memory, recomputed which costs CPU time. Intermediate options exist using check-pointing methods.

19
Data Assimilation Training Course, Reading, 5-14 May 2010 Adjoint of one instruction From the tangent linear code: dxdt(1)=-p*x(1)+p*x(2) In matrix form, it can be written as: which can easily be transposed: The corresponding code is: x(1)=x(1)-p*dxdt(1) x(2)=x(2)+p*dxdt(1) dxdt(1)=0

20
Data Assimilation Training Course, Reading, 5-14 May 2010 The Adjoint Code Property of adjoints (transposition): Application: where represents the line of the tangent linear model. The adjoint code is made of the transpose of each line of the tangent linear code in reverse order.

21
Data Assimilation Training Course, Reading, 5-14 May 2010 Adjoint of loops In the TL code for the Lorenz model we have: DO i=1,3 x(i)=x(i)+dt*dxdt(i) ENDDO which is equivalent to x(1)=x(1)+dt*dxdt(1) x(2)=x(2)+dt*dxdt(2) x(3)=x(3)+dt*dxdt(3) We can transpose and reverse the lines: dxdt(3)=dxdt(3)+dt*x(3) dxdt(2)=dxdt(2)+dt*x(2) dxdt(1)=dxdt(1)+dt*x(1) which is equivalent to DO i=3,1,-1 dxdt(i)=dxdt(i)+dt*x(i) ENDDO

22
Data Assimilation Training Course, Reading, 5-14 May 2010 Conditional statements What we want is the adjoint of the statements which were actually executed in the direct model. We need to know which branch was executed The result of the conditional statement has to be stored: it is part of the trajectory !!!

23
Data Assimilation Training Course, Reading, 5-14 May 2010 Tangent linear codeAdjoint code δx = 0δx * = 0 δx = A δy + B δzδy * = δy * + A δx * δz * = δz * + B δx * δx * = 0 δx = A δx + B δzδz * = δz * + B δx * δx * = A δx * do k = 1, N δx(k) = A δx(k 1) + B δy(k) end do do k = N, 1, 1 (Reverse the loop!) δx * (k 1) = δx * (k 1) + A δx * (k) δy * (k ) = δy * (k) + B δx * (k) δx * (k) = 0 end do if (condition) tangent linear codeif (condition) adjoint code Summary of basic rules for line-by-line adjoint coding (1) And do not forget to initialize local adjoint variables to zero ! Adjoint statements are derived from tangent linear ones in a reversed order Order of operations is important when variable is updated!

24
Data Assimilation Training Course, Reading, 5-14 May 2010 Tangent linear codeTrajectory and adjoint code if (x > x0) then δx = A δx / x x = A Log(x) end if Trajectory x store = x (storage for use in adjoint) if (x > x0) then x = A Log(x) end if Adjoint if (x store > x0) then δx * = A δx * / x store end if Summary of basic rules for line-by-line adjoint coding (2) The most common sources of error in adjoint coding are: 1)Pure coding errors 2)Forgotten initialization of local adjoint variables to zero 3)Mismatching trajectories in tangent linear and adjoint (even slightly) 4)Bad identification of trajectory updates To save memory, the trajectory can be recomputed just before the adjoint calculations.

25
Data Assimilation Training Course, Reading, 5-14 May 2010 More remarks about adjoints The adjoint always exists and it is unique, assuming spaces of finite dimension. Hence, coding the adjoint does not raise questions about its existence, only questions of technical implementation. In the meteorological literature, the term adjoint is often improperly used to denote the adjoint of the tangent linear of a non-linear operator. In reality, the adjoint can be defined for any linear operator. One must be aware that discussions about the existence of the adjoint usually address the existence of the tangent linear model. Without re-computation, the cost of the TL is usually about 1.5 times that of the non-linear code, the cost of the adjoint between 2 and 3 times. The tangent linear model is not strictly necessary to run 4D-Var (but it is in the incremental 4D-Var formulation in use operationally at ECMWF). It is also needed as an intermediate step to write the adjoint.

26
Data Assimilation Training Course, Reading, 5-14 May 2010 machine precision reached Perturbation scaling factor Test for tangent linear model

27
Data Assimilation Training Course, Reading, 5-14 May 2010 Test for adjoint model The adjoint test is truly unforgiving. If you do not have a ratio of the norm close to 1 within the precision of the machine, you know there is a bug in your adjoint. At the end of your debugging you will have a perfect adjoint (although you may still have an imperfect tangent linear!)

28
Data Assimilation Training Course, Reading, 5-14 May 2010 Test of adjoint in practice… (example from the aerosol assimilation) Compute perturbed variable (for example optical depth, ) using perturbation in input variables (for example, mixing ratio,, and humidity, ) with the tangent linear code Call adjoint routine to obtain gradients in and with respect to initial condition ( and ) from perturbation in. Compute the norm from the adjoint calculation, using unperturbed state and gradients: According to the test of adjoint NORM_TL must be equal to NORM_AD to the machine precision!

29
Data Assimilation Training Course, Reading, 5-14 May 2010 Automatic differentiation Because of the strict rules of tangent linear and adjoint coding, automatic differentiation is possible. Existing tools: TAF (TAMC), TAPENADE (Odyssée),... –Reverse the order of instructions, –Transpose instructions instantly without typos !!! –Especially good in deriving tangent linear codes! There are still unresolved issues: – It is NOT a black box tool, – Cannot handle non-differentiable instructions (TL is wrong), – Can create huge arrays to store the trajectory, – The codes often need to be cleaned-up and optimised.

30
Data Assimilation Training Course, Reading, 5-14 May 2010 Useful References Variational data assimilation: Lorenc, A., 1986, Quarterly Journal of the Royal Meteorological Society, 112, Courtier, P. et al., 1994, Quarterly Journal of the Royal Meteorological Society, 120, Rabier, F. et al., 2000, Quarterly Journal of the Royal Meteorological Society, 126, The adjoint technique: Errico, R.M., 1997, Bulletin of the American Meteorological Society, 78, Tangent-linear approximation: Errico, R.M. et al., 1993, Tellus, 45A, Errico, R.M., and K. Reader, 1999, Quarterly Journal of the Royal Meteorological Society, 125, Janisková, M. et al., 1999, Monthly Weather Review, 127, Mahfouf, J.-F., 1999, Tellus, 51A, Lorenz model: X. Y. Huang and X. Yang. Variational data assimilation with the Lorenz model. Technical Report 26, HIRLAM, April E. Lorenz. Deterministic nonperiodic flow. J. Atmos. Sci., 20: , Automatic differentiation: Giering R., Tangent Linear and Adjoint Model Compiler, Users Manual Center for Global Change Sciences, Department of Earth, Atmospheric, and PlanetaryScience,MIT,1997 Giering R. and T. Kaminski, Recipes for Adjoint Code Construction, ACM Transactions on Mathematical Software, 1998 TAMC: TAPENADE: Sensitivity studies using the adjoint technique Janiskova, M. and J.-J. Morcrette., Investigation of the sensitivity of the ECMWF radiation scheme to input parameters using adjoint technique. Quart. J. Roy. Meteor. Soc., 131,

Similar presentations

© 2017 SlidePlayer.com Inc.

All rights reserved.

Ads by Google